A "composite distribution" is a statistical technique to approximate the composition of a user group which contains several separately measured datasets. It is used in paper applications that cannot do the more accurate iterative calculations that PeopleSize Pro uses.
It is created by combining the various measured populations into a single distribution, with a single Mean and Standard Deviation. It has the advantage that once the composite distribution is created, size estimates can be made with simple calculations.
However the User Group represented by the composite distribution is unlikely to be really 'normal', but will have irregularities as the constituent populations ebb and flow in the exclusions. This can cause composite distribution estimates to have significant errors, especially at the extremes where the dispersion can be underestimated.
See also: User Groups